Abstract: A method and apparatus for parallel training of a neural network model are provided. The method includes generating a plurality of partial data sets by dividing training data for a neural network model, identifying a plurality of pipeline stages of the neural network model, determining a plurality of keeping policies, wherein each of the plurality of keeping policies corresponds to values generated by a selected pipeline stage of the plurality of the pipeline stages using a selected partial data set of the plurality of partial data sets, and training the neural network model by processing the values generated by the selected pipeline stages based on the selected partial data set according to a corresponding keeping policy for each of the plurality of keeping policies.

Abstract: A neural network operation apparatus includes: a receiver configured to receive a first input feature map; a controller configured to control multiplier-accumulators (MACs) included in a first MAC array; and a first operation engine comprising the first MAC array and configured to process the first input feature map based on the MACs of which operation states are controlled.

Abstract: A processor-implemented method with data loading includes: based on sizes of a plurality of data files in a training dataset, dividing the training dataset into a plurality of sub-sets; loading some data files in each sub-set into a plurality of processors; determining a packing combination of one or more data files loaded to processors in a same group among the plurality of processors, based on a ratio of a number of data files between the plurality of sub-sets and a batch size of distributed training; determining packed data files by packing the one or more data files according to the packing combination; and reallocating the packed data files to the processors in the same group.

Abstract: Disclosed is a skin care device using plasma. The skin care device may include a first plasma generating device; and a main body configured to supply power to the first plasma generating device and to control the first plasma generating device based on an input from a user.

Abstract: Disclosed is a plasma generator. The plasma generator may include a gripping portion including at least one interface unit configured to receive an input from a user; a head portion including a plasma generating portion configured to generate the plasma; a first cartridge configured to detachably couple at a first end of the head portion and generate the plasma over a predetermined region; and a light irradiation portion provided at a second end of the head portion.

Abstract: Disclosed is a skin care device using plasma. The skin care device may include a first plasma generating device; and a main body configured to supply power to the first plasma generating device and to control the first plasma generating device based on an input from a user.

Abstract: Disclosed is a plasma generator. The plasma generator may include a gripping portion including at least one interface unit configured to receive an input from a user; a head portion including a plasma generating portion configured to generate the plasma; a first cartridge configured to detachably couple at a first end of the head portion and generate the plasma over a predetermined region; and a light irradiation portion provided at a second end of the head portion.

Abstract: A method and apparatus for estimating execution time of a neural network are provided, the method of estimating execution time of a neural network in a multi-core accelerator, the method including generating trace information including operation timing information for each core of the multi-core accelerator, and calculating the execution time of the neural network reflecting communication overhead between cores of the multi-core accelerator and memory access time for each core of the cores, based on the trace information.

Abstract: Disclosed are a training method and apparatus for distributed training of a neural network, the training apparatus including processors configured to perform distributed training, wherein each of the processors is further configured to perform a forward direction operation for layers of the neural network, determine a loss of the neural network based on the forward direction operation, determine a local gradient for each layer of the neural network by performing a backward direction operation for the layers of the neural network based on the loss, determine whether to perform gradient clipping for a local gradient determined for a previous layer, in response to determining a local gradient for a current layer through the backward direction operation, determine an aggregated gradient based on the backward direction operation and the gradient clipping performed by each of the processors, and update parameters of the neural network based on the aggregated gradient.

Abstract: A processor-implemented method with data loading includes: dividing a training data set into a plurality of subsets based on sizes of a plurality of data files included in the training data set; loading, from each of the plurality of subsets, a portion of data files in the subset to a plurality of processors based on a proportion of a number of data files of the plurality of subsets in the subset and a batch size of distributed training; and reallocating, based on sizes of data files loaded to processors in a same group among the plurality of processors, the loaded data files to the processors in the same group.

Abstract: A neural network operation apparatus includes: a receiver configured to receive a first input feature map; a controller configured to control multiplier-accumulators (MACs) included in a first MAC array; and a first operation engine comprising the first MAC array and configured to process the first input feature map based on the MACs of which operation states are controlled.

Abstract: Provided are an imaging device production system, an imaging device inspection system, and an imaging device inspection method. The imaging device production system includes a assembling process for assembling each components constituting an imaging device, an inspection process in which the imaging device assembled in he assembling process is continuously transferred without being stopped except for emergency when two inspection tasks adjacent to and different from each other are performed, and a packaging for packaging the imaging device which passes through the inspection tasks to confirm quality of the imaging device.

Abstract: A method of manufacturing a solar cell includes preparing a base substrate having a first conductive type; diffusing an impurity having a second conductive type (opposite the first conductive type) into the base substrate to form an emitter layer having a first impurity concentration on the base substrate and a by-product layer on the emitter layer; irradiating a laser beam onto the emitter layer corresponding to a first region of the base substrate to form a front contact portion having a second impurity concentration higher than the first impurity concentration; irradiating the laser beam onto the by-product layer to remove the by-product layer corresponding to the first region; removing the by-product layer from an area outside of the first region; forming an anti-reflection layer on the base substrate; forming a front electrode on the anti-reflection layer corresponding to the first region; and forming a back electrode on the base substrate.

Abstract: In a method of manufacturing a solar cell, a first dopant layer is formed on a lower surface of a substrate and a diffusion-preventing layer is formed on an upper surface of the substrate. Then, the first dopant layer is patterned to expose portions of the lower surface of the substrate, and a second dopant layer is formed on the exposed portion of the lower surface of the substrate. A third dopant layer is formed on the diffusion-preventing layer, and the substrate is heated to diffuse dopants from the first, second, and third dopant layers into the substrate, thereby forming semiconductor areas in the substrate.

Abstract: A solar cell includes a substrate, a doped pattern, a contact layer, and an electrode. The substrate includes a first surface onto which sunlight is incident and a second surface facing the first surface. The doped pattern is formed on the second surface of the substrate and the contact layer is formed on the doped pattern. The electrode is formed on the contact layer and is electrically connected to the doped pattern. Accordingly, a contact resistance between the substrate and the electrode may be decreased, so that the doped pattern and the electrode may be uniformly formed and a power efficiency of the solar cell may be improved.

Abstract: Provided are an imaging device production system, an imaging device inspection system, and an imaging device inspection method. The imaging device production system includes a assembling process for assembling each components constituting an imaging device, an inspection process in which the imaging device assembled in he assembling process is continuously transferred without being stopped except for emergency when two inspection tasks adjacent to and different from each other are performed, and a packaging for packaging the imaging device which passes through the inspection tasks to confirm quality of the imaging device.